KTH Chemical Science
and Engineering
In-situ activated hydrogen evolution
from pH-neutral electrolytes
John Gustavsson
Doctoral Thesis
Applied Electrochemistry
School of Chemical Science and Engineering
Kungliga Tekniska Högskolan, 2012
Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan i Stockholm,
framlägges till offentlig granskning för avläggande av teknologie doktorsexamen
fredag den 15:e juni 2012, kl. 10.00 i sal E3, Lindstedtsvägen 27, Plan 5,
Kungliga Tekniska Högskolan
© John Gustavsson 2012
Printed in Sweden
TRITA-CHE Report 2012:26
ISSN 1654-1081
ISBN 978-91-7501-391-6
Abstract
The goal of this work was to better understand how molybdate and trivalent cations can be
used as additives to pH neutral electrolytes to activate the Hydrogen Evolution Reaction
(HER). Special emphasis was laid on the chlorate process and therefore also to some of the
other effects that the additives may have in that particular process.
Cathode films formed from the molybdate and trivalent cations have been investigated with
electrochemical and surface analytical methods such as polarization curves, potential sweep,
Electrochemical Impedance Spectroscopy (EIS), current efficiency measurements, Scanning
Electron Microscope (SEM), Energy-Dispersive X-ray Spectroscopy (EDS), X-ray
Photoelectron Spectroscopy (XPS), X-Ray Fluorescence (XRF) and Inductively Coupled
Plasma (ICP) analysis.
Trivalent cations and molybdate both activate the HER, although in different ways. Ligand
water bound to the trivalent cations replaces water as reactant in the HER. Since the ligand
water has a lower pKa than free water, it is more easily electrochemically deprotonated than
free water and thus catalyzes the HER. Sodium molybdate, on the other hand, is
electrochemically reduced on the cathode and form films which catalyze the HER (on cathode
materials with poor activity for HER). Molybdate forms films of molybdenum oxides on the
electrode surface, while trivalent cation additions form hydroxide films. There is a risk for
both types of films that their ohmic resistance increases and the activity of the HER decreases
during their growth. Lab-scale experiments show that for films formed from molybdate, these
negative effects become less pronounced when the molybdate concentration is reduced.
Both types of films can also increase the selectivity of the HER by hindering unwanted side
reactions, but none of them as efficiently as the toxic additive Cr(VI) used today in the
chlorate process. Trivalent cations are not soluble in chlorate electrolyte and thus not suitable
for the chlorate process, whereas molybdate, over a wide pH range, can activate the HER on
catalytically poor cathode materials such as titanium.
Keywords: molybdate, trivalent cations, electrolysis, hypochlorite reduction, films,
electrolysis, chlorate process
Sammanfattning
Målsättningen med detta doktorsarbete har varit att bättre förstå hur trivalenta katjoner och
molybdat lösta i elektrolyten kan effektivisera elektrokemisk vätgasproduktion.
Tillämpningen av dessa tillsatser i kloratprocessen och eventuella sidoeffekter har undersökts.
De filmer som bildas på katoden av tillsatserna har undersökts med både elektrokemiska och
fysikaliska ytanalysmetoder: polarisationskurvor, potentialsvep, elektrokemisk
impedansspektroskopi (EIS), strömutbytesmätningar, svepelektronmikroskopi (SEM),
energidispersiv röntgenspektroskopi (EDS), röntgenfotoelektronspektroskopi (XPS),
röntgenfluorensens (XRF) och induktivt kopplat plasma (ICP).
Både trivalenta katjoner och molybdat kan aktivera elektrokemisk vätgasutveckling, men på
olika sätt. Vatten bundet till trivalenta katjoner ersätter fritt vatten som reaktant vid
vätgasutveckling. Eftersom vatten bundet till trivalenta katjoner har lägre pKa-värde, går det
lättare att producera vätgas från dessa komplex än från fritt vatten. Natriummolybdat däremot
reduceras på katoden och bildar filmer som kan katalysera vätgasutvecklingen på
substratmaterial som har låg katalytisk aktivitet för reaktionen. Molybdat bildar
molybdenoxider på ytan medan trivalenta katjoner bildar metallhydroxider. Båda typerna av
film riskerar att bilda filmer som är resistiva och deaktiverar vätgasutvecklingen.
Laboratorieexperiment visar att problemen minskar med minskad molybdathalt.
Båda filmerna kan öka selektiviteten för vätgasutveckling genom att hindra sidoreaktioner.
Filmerna är dock inte lika effektiva som de filmer som bildas från den ohälsosamma tillsatsen
Cr(VI), vilken används i kloratprocessen idag. Trivalenta katjoner är inte lösliga i
kloratelektrolyt och är därför inte en lämplig tillsats i kloratprocessen. Molybdat har god
löslighet och kan aktivera vätgasutveckling i ett stort pH-intervall på titan och andra
substratmaterial som själva har betydlig sämre aktivitet för vätgasutveckling.
Nyckelord: molybdat, trivalenta katjoner, elektrolys, hypokloritreduktion, filmer,
kloratprocessen
Preface
This thesis comprises the present summary based primary on the following papers.
Paper I
Cathodic Reactions on an iron RDE in the presence of Y(III), Linda Nylén, John
Gustavsson and Ann Cornell, Journal of The Electrochemical Society 155 (2008)
E136-E142
Paper II
Rare earth metal salts as potential alternatives to Cr(VI) in the chlorate process,
John Gustavsson, Linda Nylén and Ann Cornell, Journal of Applied Electrochemistry 40
(2010) 1529-1536
Paper III In-situ activation of hydrogen evolution in pH neutral electrolytes by additions
of multivalent cations, John Gustavsson, Göran Lindbergh and Ann Cornell, International
Journal of Hydrogen Energy, in press (2012)
Paper IV
In-situ activated hydrogen evolution by Mo(VI)-addition to neutral and alkaline
electrolytes, John Gustavsson, Christine Hummelgård, Joakim Bäckström, Inger Odnevall
Wallinder, Seikh Mohammed Habibur Rahman, Göran Lindbergh, Sten Eriksson and Ann
Cornell, Manuscript submitted to Journal of Electrochemical Science and Engineering
Paper V
On the suppression of cathodic hypochlorite reduction by electrolyte additions
of molybdate and chromate ions, John Gustavsson, Gongzhuo Li, Christine Hummelgård,
Joakim Bäckström and Ann Cornell, Manuscript
My contributions to the different papers in this thesis
I: My experimental contributions to this paper were the potential sweeps in Figure 4. I also
contributed with experiments that initially detected nitrate reduction on the iron RDE (not
published) and was actively involved in the planning and discussions during the study.
II-IV: I performed all electrochemical measurements in these papers. I was helped with the
surface analyses, although I was active in the interpretation of the data.
V: I performed the potential sweep and the anodic polarization experiments. Gongzhuo Li
performed the current efficiency measurements as part of his Master thesis with Ann Cornell
and myself as supervisors. I was helped with the surface analyses, although I was active in the
interpretation of the data.
I wrote papers II-V with assistance of the co-authors.
TableofContents
1. Introduction ......................................................................................................................................... 1 1.1 Energy savings in industrial electrolysis ............................................................................................ 1 1.2 Hydrogen production .................................................................................................................... 2 1.3 Hydrogen evolution from pH neutral electrolytes ........................................................................ 3 1.4 The chlorate process ..................................................................................................................... 5 1.5 In‐situ additives ............................................................................................................................. 6 1.5.1 Trivalent cations as electrolyte additive ................................................................................ 7 1.5.2 Sodium molybdate as electrolyte additive ............................................................................. 8 2. Aim of the study ................................................................................................................................ 11 3. Experimental ..................................................................................................................................... 13 4. Results and discussion ....................................................................................................................... 15 4.1 Effect of electrolyte pH on hydrogen evolution .......................................................................... 15 4.2 Trivalent cations as electrolyte additives .................................................................................... 16 4.2.1 Activated hydrogen evolution by trivalent cations .............................................................. 16 4.2.2 Increased hydrogen evolution selectivity by trivalent cation additives ............................... 26 4.2.3 Practical implications of trivalent cation addition with emphasis on the chlorate process 27 4.2.4 General discussion of trivalent cation additives .................................................................. 27 4.3 Sodium molybdate as electrolyte additive .................................................................................. 28 4.3.1 Activated hydrogen evolution by sodium molybdate .......................................................... 28 4.3.2 Increased hydrogen evolution selectivity by sodium molybdate additive ........................... 32 4.3.3 Effects on the anode by sodium molybdate addition .......................................................... 36 4.3.4 Practical implications of sodium molybdate addition with emphasis on the chlorate process
....................................................................................................................................................... 37 4.4 Comparing trivalent, sodium molybdate, buffers and Cr(VI) electrolyte additives .................... 38 5. Conclusions ........................................................................................................................................ 39 5.1 Trivalent cations as additive to pH neutral electrolytes.............................................................. 39 5.2 Sodium molybdate as additive to pH neutral electrolytes .......................................................... 39 6. Future work ....................................................................................................................................... 41 7. Acknowledgements ........................................................................................................................... 41 8. References ......................................................................................................................................... 43 1.Introduction
Both for economic and environmental reasons it is important to lower the electrical energy
needed when producing hydrogen by electrolysis. In this chapter a short introduction is given
to energy savings for cathodic hydrogen evolution from neutral electrolytes. The introduction
begins with a short description of general electrolysis and then focuses more on hydrogen
evolution and the application of hydrogen evolution in the chlorate process. As the final part
of the introduction in-situ activated hydrogen evolution is presented.
1.1Energysavingsinindustrialelectrolysis
About 10% of the world’s electricity consumption is used in electrolytic processes [1]; it is
therefore important to understand the underlying factors. The electrical energy must supply
sufficient cell voltage (ΔV) to drive the electrochemical reaction and sufficient charge to
produce the desired amount of products. The minimum required cell voltage can be calculated
from the Gibbs free energy (ΔGr) for the total reaction according to equation 1, where z is the
number of electrons per molecule produced, F is the Faradays constant and ΔE is the
equilibrium voltage for the reaction.
ΔGr=-zFΔE
(1)
In reality, there are always electrical energy losses in an electrolysis cell. There are ohmic
losses due to resistances in the cell (ΔVΩ), i.e. in the electrolyte and cell hardware. The
overpotentials (η) on the cathode and anode cause further losses. Overpotential losses can be
divided into activation overpotential (ηact) and concentration polarization (ηconc). Activation
overpotential is the extra voltage required to overcome the kinetic hindrance and to drive the
reaction at the desired rate. Concentration polarization is caused by the fact that reactant
concentration can be lower and/or product concentration higher at the electrode than in the
bulk electrolyte. This leads to an increased overpotential. An equation for the cell voltage ΔV
is given as equation 2 [2].
ΔV=ΔE + ∑ |η| + ΔVΩ
(2)
The theoretical charge (Qtheoretical) required in electrolysis can be calculated by Faraday’s law:
Qtheoretical=nzF
(3)
where n is the amount of substance produced in moles.
The Faraday’s constant is 96485 C mol-1 [3] and 0.19×106 C (54 Ah) is for example required
for 1 mole of H2 produced (n=1, z=2). Large-scale electrolytic production of hydrogen
therefore requires massive amounts of charge. If the current efficiency (ϕ) for the reaction is
below 100%, extra charge is required due to side reactions:
Qreal=nzF/ϕ
(4)
1 The minimum theoretical electrical energy (Etheoretical) required can be calculated using the
equilibrium voltage and the theoretical charge (equation 5). In reality the cell voltage and the
applied charge (Qreal) should be used (equation 6). The efficiency (ε) of the electrolysis can be
calculated as the ratio of the theoretical electric energy demand divided by the real electric
energy (Ereal) demand (equation 7).
Etheoretical=ΔE×Qtheoretical
(5)
Ereal=ΔV×Qreal
(6)
ε=Etheoretical/Ereal
(7)
1.2Hydrogenproduction
Electrochemical production of hydrogen takes place as main or side reaction in almost all
industrial electrolytic processes with aqueous electrolytes. The hydrogen-producing electrode
is often the auxiliary electrode such as in for example the chloralkali and chlorate processes.
In water electrolysis, microbial electrolysis cells (MEC) and water splitting solar cells
hydrogen production is the main objective. Hydrogen production is an undesired reaction in
processes such as in electrodeposition of metals and as the cathode reaction in corrosion [4].
In acidic electrolytes, hydrogen is produced by hydronium ion reduction (reaction 8) while in
alkaline electrolytes, hydrogen is produced by water reduction (reaction 9).
2H3O+ + 2e- 2H2O + H2
(8)
2H2O + 2e-  2OH- + H2
(9)
The choice of cathode material will affect the kinetics of the hydrogen evolution reaction
(HER). Two important factors at technical current densities are the exchange current density
(i0) and the Tafel slope (B). The relationship between the activation polarization (ηact) and the
current density can be expressed by the Tafel equation [5]:
ηact=A+B×log i
(10)
where, for a cathodic reaction as the HER
A=ln10×RT/(αc×F)×log i0
(11)
B=-ln10×RT/(αc×F)
(12)
where i is the current density, R the universal gas constant, T the temperature, αc the cathodic
transfer coefficient, and A and B constants.
For hydrogen evolution, i0 varies several orders of magnitudes for different metals. For lead,
i0 is 10-14 A m-2 and for rhodium, i0 is 3×10-3 A m-2 [6].
Changing cathode materials in an industrial process can be expensive and difficult. Catalytic
noble metals are expensive and often easily deactivated by electrolyte impurities. The
2 requirements on the material are quite demanding. It should be a good electron conductor,
active for the HER, stable, and inexpensive, have a long life time, be easy to machine and not
interfere with the process by, for example, corroding into products that catalyze unwanted
reactions.
The geometrical dimensions of the electrolysis cell will also affect the cell voltage. A short
distance between the electrodes reduces the potential drop in the electrolyte. Cells that are
built to maintain a high convection of the electrolyte also decrease problems with mass
transport limitations and therefore lower the concentration overpotentials. Hydrogen bubbles
formed can increase the convection at the cathode by gas lift of the electrolyte and may also
increase the ohmic drop in the cell gap, shield the electrode surface and erode coatings on the
cathode. Hydrogen diffusing into the cathode material can also form hydrides which may
shorten the lifetime of the cathode.
The electrode area should be large to decrease the required current density for producing a
product at a given rate. Lower current density leads to lower potential drop in the electrolyte
and lower overpotentials. Large electrodes, though, increases the capital cost for an
electrolysis plant and there is a trade-off between large electrodes and low capital costs. Often
high capital costs require that the cell be used at a high current density to be economical
feasible. Porous electrodes and finely divided particles of the catalyst in a matrix can be used
to increase the real surface area of the electrodes and thereby decrease the real current density.
1.3HydrogenevolutionfrompHneutralelectrolytes
In processes where strongly alkaline or acidic conditions are not possible, there will be
concentration overpotentials. As seen in Fig. 1, reactions 8-9 will increase the pH close to the
cathode while for example oxygen evolution (reactions 13-14) will decrease the pH close to
the anode. For reactions 8 and 9 this will lead to increased concentration overpotentials.
Fig. 1: Schematic pH profile for water electrolysis with neutral bulk electrolyte pH.
3 2H2O  O2 + 4H+ + 4e-
(13)
4OH-  O2 + 2H2O + 4e-
(14)
Whenever hydrogen production from slightly acidic or neutral electrolytes is desired either as
an auxiliary reaction or as the main purpose of the process, concentration overpotentials may
be a problem as regards achieving a high energy efficiency. For water splitting solar cells
some catalysts may work optimally around neutral pH [7]. In MEC, around neutral pH is the
optimal pH range for the microorganisms oxidizing organic matter on the anode [8-10].
Adding buffer to the electrolyte and increasing convection decrease the overpotential and thus
increase the energy efficiency [9]. Optimal pKa value of the buffer is close to the pH value of
the bulk electrolyte [9].
The equilibrium potentials for reactions 8 and 9 are the same, but their relative rates depend
on the conditions. Typically, in electrolysis in near pH neutral electrolyte, hydronium ion
reduction (reaction 8) is the fastest at low current densities while water reduction (reaction 9)
dominates at high current densities [11,12]. During the course of a polarization curve, the pH
at the cathode surface increases with increasing current density and, when the transport of
H3O+ to the surface becomes rate limited, the reaction changes from 8 to 9. The magnitude of
this limiting current density depends on factors as buffer concentration and mass transport
conditions [11].
A more general way of writing equations 8 and 9 is
2H+B + 2e-  H2 + 2B
(15)
where B is a Brønsted base such as H2O (reaction 8), and OH- (reaction 9). In principle, any
hydrogen-containing specie that has a lower pKa than water can act as proton donor (assuming
no kinetic limitation) in a catalyzed HER. The pKa values for water and hydronium ions are
15.73 and -1.74, respectively [13]. The large difference in pKa values between water and
hydronium ions explains why hydrogen is preferably evolved from hydronium ions rather
than from water molecules.
Two mechanisms for reaction 15 that differ in principle have been suggested: i) with H+B
dissociating to H+ + B followed by reaction 8, and ii) direct electrochemical deprotonation of
H+B [14,15]. The direct deprotonation has been concluded for example for phosphate ions
[14,16], phosphoric acid anions[17], acetic acid [15], citric acid [18], pyrophosphoric acid
[19] and ammonium ions [20].
4 1.4Thechlorateprocess
One major industrial process where hydrogen is evolved from neutral electrolyte is the
chlorate process. Sodium chlorate is produced in an electrical energy intensive process where
sodium chloride is oxidized into sodium chlorate and with hydrogen gas as a product at the
cathode (reaction 16).
(16)
NaCl + 3H2O NaClO3 + 3H2
®
Dimensionally stable anodes (DSA ) are used industrially as anodes while the cathodes are
normally made from low-alloyed carbon steel or titanium. Steel cathodes corrode, which not
only shortens cathode lifetime - corrosion products that form contaminate the product and can
cause short circuiting in the cells. Although, there are also positive side effects of corrosion,
such as that the electrode area may increase and thereby decrease the overpotential for the
HER. Steel cathodes have overpotentials of around 800 mV for hydrogen evolution
(reaction 9) [21].
Titanium cathodes are more corrosion-resistant than steel cathodes. Drawbacks are that
titanium hydride forms over time and Ti shows even higher overpotentials than steel. The
electric energy used for electrolysis accounts for up to 80% of the variable production cost of
chlorate, and with rising costs for electricity, a reduction of the cathode overpotential is of
utmost importance. Another major concern is the search for an alternative to the carcinogenic
Cr(VI) electrolyte additive. At present, Cr(VI) is necessary to obtain high current efficiency of
the process. During operation, Cr(III) is electrodeposited on the cathode, forming a film of
chromium hydroxide that hinders the electroreduction of chlorate (reaction 17) and
hypochlorite (HClO + ClO-) (reaction 18), an intermediate in the chlorate process [22]. The
chromium hydroxide film is only a few nm thick [23] and can inhibit its own growth [22]. It
also affects the kinetics of HER so it decreases the differences in overpotentials for HER
between different electrode materials [24]. Chromate also functions as a buffer to keep the
electrolyte pH at 5.9-6.7, which is optimal for the process [25]. As can be seen in reaction 16,
the overall production of chlorate is in principle pH neutral although some acid has to be
added since some Cl2 and HClO escape with the cell gas.
ClO3- + 3H2O +6e- Cl- + 6OHClO- + H2O +2e- Cl- +2OH-
(17)
(18)
It should be noted that even if the electrolysis cell is the most electrical energy consuming
part of the chlorate process, and the focus of this work, there are several other process steps in
the chlorate process (Fig. 2). The sodium chloride salt has to be purified in several steps
before entering the process. Due to environmental concerns, the process is an almost closed
system and impurities entering the process will be enriched in the chlorate electrolyte and
may deposit on the cathode surfaces. The product, sodium chlorate, is separated in a
crystallization step at an increased pH to avoid corrosion of the crystallizer. Some acid and
base are therefore needed to adjust the pH up before crystallization and then down before
entering the electrolysis cells again.
5 Fig. 2: Schematic process flow chart of the chlorate process (Courtesy of Akzo Nobel)
1.5In‐situadditives
Activation of the HER by in-situ additives has several advantages over replacing electrode
materials or designing porous electrodes. In-situ additives are relatively simple and quick to
implement in an existing process. In-situ addition to an electrochemical process can be
performed in several different ways. For example, it can be done such that the process
electrolyte is temporary replaced with an electrolyte containing the additives, the electrodes
are treated in a separate electrolyte bath before being transferred to the process, or the
additives are added directly into the process electrolyte. It can be even simpler and faster to
add the additives directly to the process electrolyte than other alternatives which involve
stopping the process to replace electrolyte or electrodes. Furthermore, if the additives are
present in the process electrolyte, the effect of the additives can be continuously renewed.
When searching for suitable electrolyte additives, only a minor subset of all known elements
and compounds are of interest for activated HER from aqueous electrolytes. Compounds may
also have negative effects as additives such as deactivating the HER, contaminating the
product or in other ways interfering with the process. Since this work was performed with an
emphasis on the chlorate process, some specific examples are of interest. Addition of organic
compounds to the chlorate process will be difficult to gain acceptance for, due to the risk of
forming environmentally harmful chlorinated organic compounds. Chlorate salts may also
form explosive mixtures with organic compounds. Elements such as cobalt and nickel are not
desirable since they may form ions that decompose hypochlorite [21], which is an
intermediate in the process. The sodium chloride added to the electrolyte has to be purified to
remove undesired compounds, see Fig. 2.
Components in a process electrolyte can for example activate the HER in the following ways:
form a catalytically active film on the cathode surface, clean the cathode surface from
deactivating deposits/films [26], form a film which increases the real surface area of the
6 cathode, increase the hydronium ion concentration at the cathode surface (buffers) or replace
water as reactant (compounds containing hydrogen atoms with lower pKa than water).
This work focuses on two types of film-forming additives that in different systems show
proven abilities to both activate the HER and inhibit side reactions; trivalent cations and
Mo(VI), i.e. sodium molybate.
1.5.1Trivalentcationsaselectrolyteadditive
Generally, in acidic solutions trivalent cations are soluble, and in neutral to alkaline
electrolyte the ions precipitate as hydroxides or oxides. In Fig. 3, it can be seen that Y(III)
precipitates around neutral pH. The exact pH of precipitation will depend on factors as the
trivalent cation ion concentration, the ionic strength of the solution and if there are anions
forming complexes.
-1.0
-1.5
Y 3+
Y (O H )3
Log Conc. / M
-2.0
-2.5
-3.0
Y O H 2+
-3.5
-4.0
0
2
4
6
8
pH
10
12
14
Fig. 3: Log concentration vs. pH diagram generated with the Medusa Software [26]. The total Y(III)
concentration was set to 15 mM.
The solubility of the ions depends on the pH (Fig. 4). In neutral solutions, Al(III) and Sc(III)
have low solubility, while Sm(III), Y(III) and La(III) will precipitate through a slight pH
increase.
At alkaline pH, the solubility is a little higher than the minimum value due to M(OH)4formation. Sc(III), Sm(III), Y(III), and La(III) nonetheless form precipitations at pH 14
if 15 mM is added to the electrolyte.
7 0
-2
La(III)
II)
Y(I
III)
Sm(
Log Solubl. / M
Sc(III)
-4
-6
-8
Al(III)
-10
0
2
4
6
8
10
pH
12
14
Fig. 4: Solubility diagram generated with the Medusa Software [26].The total M(III) concentration
was set to 15 mM for each metal ion.
The trivalent cations Y3+, Sc3+, Sm3+ and La3+ are rare earth metals (REM). The addition of
Y(III) to a pH neutral electrolyte has been found to catalyze the HER [28-31]. It can be
assumed that Y(III) is not reduced, as the E0Y(III)/Y = -2.37 V, and it has been suggested that
water in Y(III) aqua complexes are the HER reactants [28-31]. A film of Y(OH)3 precipitates
during electrolysis in the alkaline layer on the cathode surface (see Fig. 3 and Fig. 4) and this
film can hinder the reduction of oxygen [28], a cathode reaction in electrochemical corrosion.
This dual function of Y(III) addition, to activate the HER and simultaneously hinder cathodic
side reactions, makes it interesting for processes where hydrogen is evolved in near pH
neutral solutions. The electric energy consumption to produce a given amount of hydrogen is
proportional to ΔV/. Decreasing the overpotential for HER, and thereby the cell voltage, or
increasing the current efficiency will both reduce the electric energy consumption. For the
Y(III)-activated HER, a reaction similar to 15 has been suggested, where the reactant is an
Y(III) aqua complex, see reaction 19 [28-31].
Y(H2O)x3+ + 3e− → Y(H2O)x−33+ + 3/2H2 + 3OH−
(19)
Better understanding of how Y(III) and other trivalent cations can be used to activate the HER
and prevent side reactions are of great interest. It is also important to consider factors that may
prevent practical use in applications such as the chlorate process. To our knowledge, such
studies with an emphasis on the chlorate process cannot be found in the open literature.
1.5.2Sodiummolybdateaselectrolyteadditive
Mo(VI) has shown some promising features as in-situ activator for hydrogen evolution, both
in chlorate electrolyte [32,33] and in strongly alkaline solutions [34-39]. What makes it even
more interesting is the ability of in-situ formed Mo-containing films to suppress cathodic
oxygen reduction [40] (and thereby possibly also the reduction of other species such as
8 hypochlorite). Mo(VI) is regarded as an environmentally friendly alternative to Cr(VI) in
corrosion protection [40].
Li et al. [32] have investigated the possibility to replace Cr(VI) by Mo(VI) in the chlorate
electrolyte. They concluded that addition of Mo(VI) decreased the overpotential for hydrogen
evolution by 100-130 mV on steel cathodes, that the two compounds had comparable buffer
capacity, and that the Mo(VI) addition increased the levels of unwanted oxygen in the cell
gas. Later studies [33] indicated that the effect on oxygen levels depends on the Mo(VI)
concentration, and that low concentrations, 1-10 mg dm-3 MoO3 (7-70 μM Mo(VI)) can
activate the HER with no increased oxygen production. Such low Mo(VI) levels require, in
the absence of Cr(VI), an additional buffer as phosphate to stabilize the electrolyte pH.
Under strongly alkaline conditions, 30 wt% KOH at 70oC [34-39], it was found that films
containing molybdenum and iron, the latter present as an impurity in the electrolyte,
co-deposit on the metal cathodes and result in hydrogen overpotentials at -1 kA m-2 that are
virtually independent of the substrate material. Typically, after 17 hours of polarization in the
presence of 4 mM Mo(VI) the observed overpotentials for hydrogen evolution on metal
substrates of Co, Cu, Fe, Mo, Nb, Ni, Pd, Pt, V, W and Zr varied within 40 mV compared to a
650 mV variation in the absence of Mo(VI) [37]. In the absence of electrolyte impurities such
as iron, no molybdenum-containing deposits were found [39]. Elemental molybdenum cannot
be electrodeposited from aqueous electrolytes, but as alloys with iron group metals [41] or as
molybdenum oxides [40].
The ability of electrolyte additions of Mo(VI) to suppress cathodic oxygen reduction has been
investigated on copper [40]. The inhibiting effect was most efficient at pH 8.2 and became
inefficient at pH 11, probably due to differences in the cathodic surface films formed in the
electrolytes of varying pH. In particular the oxide MoO2, a probable constituent of a cathode
film formed by electroreduction of Mo(VI), has a limited stability in alkaline solution [40,42].
Sodium molybdate and other Mo(VI)-compounds have been considered as additives to the
chlorate process in the literature. Further clarification is needed on the composition of formed
cathode films and in what way electrolyte parameters affect their growth and function.
Polarization curves where the potentials for the HER can be recorded in combination with
surface analyses and current efficiency measurements can be used to gain a clearer picture of
the function. Practical implications of Mo(VI)-addition to pH neutral electrolytes, with
emphasis on the chlorate process, are very important to consider before larger scale
application.
9 10 2.Aimofthestudy
My aim with this work was to better understand different methods to activate the cathodic
hydrogen evolution reaction from pH neutral electrolytes by in-situ additions. Particular
emphasis is laid on the chlorate process and therefore also on some of the other effects that
additions may have in that process.
I have chosen to investigate activated HER by in-situ addition of either trivalent cations or
Mo(VI) to the electrolyte. These additives form films on the cathode, which influence the
selectivity and activity for the HER. Further understanding of how and why the HER is
activated was one of the objectives. Another objective was to investigate how the formed
films affected the current efficiency for the HER and thus the overall energy efficiency.
Studying the long term effects of the additives is beyond the scope of this thesis. Most studies
were performed in a model NaCl electrolyte without sodium chlorate to avoid the risk of
cathodic chlorate reduction interfering with the results.
11 12 3.Experimental
This section is intended as a short presentation of the experimental techniques used. For a
more detailed description, the reader is referred to the individual papers.
In this work several different electrochemical methods and surface analytical techniques have
been used. For all electrochemical experiments except the current efficiency measurements,
Rotating Disc Electrodes (RDEs) were used to have a well-defined mass transport. For these
rotating disc experiments a jacketed glass vessel was used as the cell and the temperature was
controlled by a water bath. As a counter electrode a platinum grid was used and the reference
electrode was connected to the cell by an ion bridge connected to a Luggin capillary. The tip
of the Luggin capillary was placed a few mm under the working electrode. To gain more
stable potential readings, a platinum wire in contact with the electrolyte was connected by a
capacitor to the reference electrode. For galvanostatic polarization and polarization curves the
ohmic potential drop was corrected with a current interrupt technique [43]. Since the
electrolyte composition, temperature, current density and exact position of the electrodes
varied, all factors affecting the ohmic drop, the corrected potentials gave more reliable values
than non-corrected potentials. It is also easier to compare the potentials between different
works if the potentials are corrected. Potential sweep and cyclic voltammetry experiments
were used to investigate limiting currents for side reactions, oxidation of surface films and to
get an overview over the systems before further experiments. Electrochemical impedance
spectroscopy (EIS) was used to understand trends in charge transfer and electrolyte resistance.
To measure the current efficiency (CE) for the HER a divided cell [44] was used where the
cathode and anode compartments were separated by either a membrane or a diaphragm. The
working electrode in the cathode compartment was stationary, although some convection was
achieved by using a magnetic stirrer. The divided cell was immersed into a water bath for
temperature control.
Cathode films formed from the molybdate and trivalent cations have been investigated with
surface analytical techniques: Scanning Electron Microscope (SEM), Energy-Dispersive Xray Spectroscopy (EDS), X-ray Photoelectron Spectroscopy (XPS), X-Ray Fluorescence
(XRF) and Inductively Coupled Plasma (ICP) analysis. SEM was used to get images of the
formed films. EDS gave the elemental composition of the film in selected points/areas
observed with SEM. For molybdenum containing films too thin to be detected by EDS or
XRF, XPS was used. XPS is a very surface sensitive technique with a penetration depth of
only 5-10 nm, and can be used to detect the elemental composition and also individual
oxidation state of different elements in the outermost layer of the film. XRF is a relatively
quick and easy technique, and was used to get a relative comparison on how much
molybdenum compounds that could be found on electrodes polarized under different
conditions.
13 14 4.Resultsanddiscussion
In this chapter a summary of the most important experimental findings are given.
4.1EffectofelectrolytepHonhydrogenevolution
Hydrogen may be evolved through reduction of hydronium ions or of water, depending on
pH. When the bulk pH is low enough, the reduction of hydronium ions (Reaction 8) is
kinetically favored on iron [45]. The reaction may be mass transport limited at higher current
densities, in which case water reduction (Reaction 9) dominates. Whereas proton reduction is
pH-dependent, reduction of water is pH-independent whenever the reverse reaction of
hydrogen oxidation is insignificant. This may be seen in the polarization curves on iron in
NaCl electrolyte of varying pH presented in Fig. 5. The limiting currents due to poor transport
of H3O+ are proportional to the H3O+ concentration for pH 2 to pH 4, in agreement with the
Levich equation [46]. Hypochlorite was present in the electrolyte since it had formed at the
anode during the polarization. For pH 6, the limiting current observed for the polarization
curve recorded in the anodic direction is due to hypochlorite reduction. The absence of a clear
limiting current for the pH 6 curve recorded in the opposite direction, i.e. cathodic direction,
is due to the slow buildup of hypochlorite at lower current densities. At the lowest current
densities (below 2 A m−2) the polarization curves in Fig. 5 are almost horizontal when
recorded in the anodic direction, probably reaching a corrosion potential.
Fig. 5: Polarization curves on an iron RDE (3000 rpm) in 0.5 M NaCl at different pH values. The
arrows indicate whether the polarization curves were recorded in anodic () or cathodic ()
direction. (Paper 1)
It has been shown in the literature that addition of a buffer can increase the limiting current
for the HER. For example, a chromate buffer has been shown to increase the limiting current
for H3O+ reduction by both experimental data and modeling [11]. Increased mass-transport
15 and increased buffer concentration can together or individually increase the limiting current
for the HER [9,11].
4.2Trivalentcationsaselectrolyteadditives
Two different aspects of trivalent cation additives have been investigated in this section;
activation of the HER and inhibition of hypochlorite reduction.
4.2.1Activatedhydrogenevolutionbytrivalentcations
General features of the activation
As seen in Fig. 6, trivalent cations can activate the HER. Fig. 6 has been divided into three
regions. In region I, the HER activity was similar in both the presence and absence of Al(III).
An activation of the HER can be seen in region II, appearing abruptly at ≈-100 A m-2. The
current density at the transition between regions I and II will further on be called a limiting
current density. In region III, there was a limiting current for hydronium ion reduction in the
absence of Al(III). This limiting current also contained a minor contribution from the
reduction of hypochlorite formed by reactions on the counter electrode. The potential reached
a stable potential plateau at about -0.55 V, likely a corrosion potential. In the presence of
Al(III) there was no limiting current or corrosion potential in region III. The addition of
Al(III) probably formed a cathode film of Al(OH)3 which suppressed the corrosion and
inhibited the reduction of hydronium ions and hypochlorite. Regions similar to those in Fig. 6
were also found for additions of Y(III), Sm(III), and La(III).
Fig. 6: Galvanostatic polarization curves on an iron-RDE at 3000 rpm in 0.5 M NaCl with and without
10 mM AlCl3 at pH 3.5 and 25°C. Polarization curves from high to low cathodic currents. (Not in
appended papers) Hysteresis effect on the activation
Gold is less sensitive to corrosion than iron and therefore a more suitable electrode material
when investigating cation effects at low cathodic current densities that correspond to a
potential region where iron may corrode. Polarization curves measured in anodic direction
16 (from high to low current densities) and in cathodic direction on gold with Y(III) electrolyte
addition are compared in Fig. 7. The value of the limiting current density was strongly
dependent on the direction of the scan, so that pre-polarization at high current density (anodic
direction) resulted in a lower limiting current density and a narrower transition region (more
distinct limiting current). It is probable that more yttrium hydroxide would have deposited
when reaching the limiting current region in a scan in anodic direction than in a scan in
cathodic direction. The formed film probably partially blocked the electrode surface and
lowered the effective diffusion coefficients of species passing through the film.
Fig. 7: Galvanostatic polarization curves on a gold-RDE at 3000 rpm in 0.5 M NaClO4 with 10 mM
Y(ClO4)3 at pH 6.5 and 25°C. (Paper 3)
Mass transport dependence on the activation
In Fig. 8, polarization curves on gold in 0.5 M NaClO4 at close to neutral pH, in the absence
and in the presence of Y(III), are shown. They were all recorded in cathodic direction, from
low to high cathodic current densities. The electrode reaction under these conditions is
hydrogen evolution from water reduction, and with no Y(III) present a straight Tafel slope is
observed. In the presence of Y(III) the curves have a different shape; in the low current
density region, there is a parallel shift of the polarization curve to more positive potentials
corresponding to an activation of hydrogen evolution of 100-150 mV. At high cathodic
current densities, the electrode potential does not depend much on Y(III) addition, although a
small deactivation can be observed in the presence of Y(III). The magnitude of the limiting
current depends on the concentration of Y(III) in the electrolyte. The higher the concentration
of Y(III), the higher the cathodic limiting current density. Limiting currents calculated with
the Levich equation (marked by vertical lines in Fig. 8) agrees well with observed limiting
currents. In Paper 3 it was also found that higher rotation rates of the RDEs correlates to
higher limiting currents.
17 Fig. 8: Polarization curves recorded in the cathodic direction, Au-RDE 3000 rpm. The electrolyte was
0.5 M NaClO4 at pH 6.5 and 25°C with varying Y(III) concentrations. (Paper 3)
Deposition of yttrium hydroxide
The growth rate of yttrium hydroxide films on Ni electrodes in Y(NO3)3 solutions has earlier
been found to increase with increasing cathodic current density, polarization time and
Y(NO3)3 concentration [47]. Two of these factors; current density and polarization time, were
varied for ICP measurements of deposited Y in the present study (The method is described in
Paper 3). Galvanostatic polarization measurements were conducted on iron using
10 mM YCl3 at three different current densities; -79, -158 and -3154 A m-2. Based on the
amount of Y analysed and assuming it corresponded to Y(OH)3 on the surface, the efficiency
was defined as how large a proportion of produced hydroxide ions was deposited and attached
as yttrium hydroxide. The amount of produced hydroxide ions was calculated using Faradays
law assuming 100% current efficiency for hydrogen evolution. In Fig. 9, a polarization curve
is shown for hydrogen evolution on an iron electrode in an electrolyte of the same
composition as in the leaching experiments. As can be seen, the three current densities
chosen, -79, -158 and -3154 A m-2, correspond to the activated region, the transition region
(limiting current density) and the non-catalysed region, respectively. The electrode potential
at -158 A m-2 gradually decreased during the polarization, as indicated by the arrow in Fig. 9,
whereas the potentials at -79 A m-2 and -3154 A m-2 were more stable.
18 Fig. 9: Amounts of Y deposited on iron-RDE at 3000 rpm after 5 min of polarization at -79, -158
and -3154 A m-2 in 0.5 M NaCl with 10 mM YCl3 at pH 5.6 and 25°C. Potential reading during the last
4 min are shown and a polarization curve recorded in anodic direction is included for illustrative
purposes. (Paper 3) A significant amount of yttrium was found, both at current densities within the transition
region (860 μg Y cm-2) and at cathodic current densities below the transition region
(500 μg Y cm-2). Calculating efficiencies for Y(OH)3 deposition based on the amount of
hydroxide ions produced in reaction 19 gives deposition efficiencies of >59% and >69%,
respectively, which indicates that the film was strongly attached to the surface despite the
high electrode rotation rate of 3000 rpm. At the highest cathodic current densities, the amount
of Y(III) on the electrode was lower (140 μg Y cm-2, corresponding to >0.48% efficiency for
Y(OH)3 deposition), which may be connected to vigorous hydrogen evolution disrupting the
film. Assuming that the calculated limiting current density of -218 A m-2 was related to the
transport of yttrium aqua complexes reacting in a three-electron reaction, the efficiency for
yttrium hydroxide deposition based on the transport of yttrium was about 7%, so the film
appears to be destroyed by vigorous gas evolution.
Leaching test experiments were also made in chloride free electrolyte, 0.5 M NaClO4, for iron
and gold electrodes at -70 A m-2. Samples were analysed after varying periods of polarization;
5, 15 and 30 minutes. Results in Table 1 show that the deposited amount is much less after
5 minutes, compared to the results in the chloride-containing electrolyte above at -79 A m-2.
The amount of yttrium on the gold electrode increased continuously from about 290 g cm-2
after 5 minutes to 1600 g cm-2 after 30 minutes of polarization, whereas on iron the amount
of yttrium seemed to stabilise after about 15 minutes.
19 Table 1: Surface concentration of Y on the rotating disc electrode according to ICP analysis. The
efficiency for Y(OH)3 deposition based on the calculated amount of hydroxide produced is shown in
curled brackets. (Paper 3)
Time at 3000 rpm (min)*
Au-RDE
mY (μg Y cm-2)
Fe-RDE
mY (μg Y cm-2)
5
15
30
290 {>45%}
870 {>45%}
1600 {>41%}
260 {>40%}
800 {>41%}
750 {>19%}
*
Polarized at -70 A m-2 in 0.5 M NaClO4 with 15 mM Y(ClO4)3 at pH 6.7
Electrochemical impedance spectroscopy
Electrochemical impedance spectroscopy was used for further investigation of the kinetics in
the presence of Y(III). At -1.2 V vs. Ag/AgCl and 15 mM Y(III) on a gold-RDE at 3000 rpm,
the current density is in a region where hydrogen evolution is activated, see Fig. 7. Impedance
spectra under similar conditions, in both the presence and absence of Y(III) in the electrolyte,
are given as Nyquist plots in Fig. 10. Addition of 15 mM Y(III) drastically changed the EIS
response; first, a significant decrease in charge transfer resistance for the HER correlating
well to the observed activation in the polarization curves and second, the appearance of a
linear region at low frequencies. Also, without Y(III) there is at least one, possibly two,
merged suppressed half-circles, whereas with Y(III) there is a single suppressed half-circle
indicating a rate-determining kinetic step. The linear region at low frequencies may indicate a
resistance to mass transport, a Warburg impedance.
70
-Im(Z)/Ωcm
2
60
50
40
Without Y(III)
30
10 Hz
20
10
0
0
1 kHz
100 HZ
1 Hz
15 mM Y(III)
0.1 HZ
10
20
30
40
0.1 Hz
50
60
2
70
80
90
Re(Z)/Ωcm
Fig. 10: EIS (high to low frequencies) was recorded at -1.2 V vs. Ag/AgCl using an Au-RDE at
3000 rpm. The electrolyte was 0.5 M NaClO4 at pH 6.7, 25°C with and without 15 mM Y(ClO4)3. 20 Impedance spectra in the presence of Y(III) were repeatedly recorded to determine if
stationary conditions were obtained. Spectra were recorded from high to low frequencies (Fig.
11). There was an increase in charge transfer resistance as well as in electrolyte resistance
with time, thus non-stationary conditions. Increases in electrolyte resistance in combination
with the results in Table 1 suggest a build-up of a resistive film. Increased yttrium
concentration on the electrode surface depended on the build-up of the film.
This deactivation with time appeared to increase the slope of the linear region to <1 when the
spectra were recorded from high to low frequencies. Under stationary conditions, the linear
region would most likely have a slope of 1, which is the slope of a true Warburg impedance,
indicating a mass transport resistance. The deviation from the expected slope is thus most
probably due to the non-stationary conditions.
16
14
-Im(Z)/Ωcm2
12
10
Consecutive measurements
8
6
100 Hz
4
1 kHz
2
0
0
0.1 Hz
2
4
6
8
10
12
Re(Z)/Ωcm2
14
16
18
20
Fig. 11: Multiple EIS (High to low frequencies) recorded in series at -1.2 V vs. Ag/AgCl using an
Au-RDE at 3000 rpm. The electrolyte was 0.5 M NaClO4 with 15 mM Y(ClO4)3 at pH 6.7 and 25°C.
(Paper 3)
Correlation between acidity and activation
If it is ligand water from the hydration shell of metal ions that is reduced during activated
HER (as for Y(III) in reaction 19), a correlation between the acidity of the metal ion and the
polarization of the ligand water is expected. Acidic metal ions can act as Lewis acids, pulling
electrons from oxygen in the ligand water and thus weakening the HO-H bond [51]. To study
this further the following experiment was conducted; an iron-RDE was galvanostatically
polarized with and without additions of metal salts to the electrolyte. The concentrations were
the same for all salts, 15 mM, and the current density was -70 A m-2, which is below the
limiting current density for the chosen metal ion concentration and thus in a region where an
activation of hydrogen evolution is expected. Three example curves are given in Fig. 12– with
no addition, with Sc(ClO4)3 and with Al(ClO4)3. As can be seen, Sc(III) addition resulted in an
21 activation of about 350 mV. The electrode potential after Al(III) addition was initially higher
than -0.9 V, but decreased during the polarization, a deactivation probably due to precipitation
of Al(OH)3(s) on the cathode surface, reducing the active electrode area for hydrogen
evolution [48]. Al(III) showed the most pronounced deactivation with time when comparing
metal ion additions of Sc(III), Y(III), La(III), Sm(III) and Al(III). Since aluminum, as opposed
to yttrium, samarium, lanthanum, and scandium, is not a rare earth metal, it is possible that its
hydroxide has different properties than the REM hydroxides, and that these properties
promote blocking of the electrode surface. The Al-O bond can have a more covalent nature
than the REM-O bond in the formed precipitates. Possibly, networks of covalent Al-O bonds
formed and more efficiently blocked the surface. Note that iron is cathodically protected at
these potentials, so there was no corrosion of the iron electrode.
Fig. 12: Galvanostatic polarization at -70 A m-2 using a Fe-RDE at 3000 rpm. The electrolyte was
0.5 M NaClO4 at pH 3.5 and 25°C, with and without additions of 15 mM Al(ClO4)3 and
15 mM Sc(ClO4)3. (Paper 3) In Fig. 13, the y-axis represents the degree of activation by metal ion addition, which is the
difference between the electrode potentials measured in the absence and in the presence,
respectively, of the different metal ions. The x-axis represents the pKa-values from the
literature [49] for the first hydrolysis step of the respective metal ions, reaction 20:
M3+(aq) + H2O  MOH2+(aq) + H+(aq)
(20)
As can be seen, there is a clear correlation between the acidity of the metal ion and the
magnitude of the activation, as more acidic metal ions give a larger activation. Deviations
from this correlation for Al(III) addition probably depend on the presence of a suppressing
film. Correcting for this deactivation by adjusting the marker for Al(III) in Fig. 13, see the
dotted circle representing activation after 1 minute of polarization, gives a clear correlation
also including Al(III).
22 For comparison, a di-valent cation, Mg2+, was also investigated. Addition of Mg(II)
deactivated the HER for the whole 15-min polarization, also here probably caused by a
cathode film of Mg(OH)2 [50]. A linear regression, where the corrected point for Al(III) is
used and the point for Mg is excluded, gives a slope of -79 mV/pKa.
Polarization curves, not shown here, on iron in 0.5 M NaCl with addition of La(III), Sm(III)
and Al(III) were all similar to those for Y(III) addition. The limiting current density was about
the same for all curves when measured at the same electrode rotation rate, scan direction and
metal ion concentration, whereas the electrode potential in the activated region varied as
expected from the results shown in Fig. 13.
Fig. 13: Degree of activation after 10 min of galvanostatic polarization at -70 A m-2 with 15 mM
Mz+(aq) compared to the potential at the same pH without any additions. Fe-RDE at 3000 rpm in
0.5 M NaClO4 and 25°C, with and without additions. The given pKa values correspond to the
reactions:
M3+(aq) +H2O M(OH)2+(aq) + H+(aq) and
Mg2+(aq) + H2O Mg(OH)+(aq) + H+(aq).
Corrected Al(III) after 1 min of polarization. (Paper 3)
Industrial conditions
To approach more industrially relevant conditions first the effect of increased temperature
was studied in 0.5 M NaCl, and later the chloride concentration was raised to 5 M NaCl.
Polarization curves in 0.5 M NaCl with and without 10 mM YCl3 at varying temperature were
recorded, and the curves made in the presence of YCl3 are given in Fig. 14. At cathodic
current densities higher than 100 A m-2 there was no major effect of adding YCl3 to the
electrolyte, whereas at lower current densities the addition made the potential shift in anodic
direction. This shift is related to an activation of the HER by Y(III) ions, as Y(III) is not redox
active under these conditions. This catalytic effect clearly decreased with increasing
23 temperature, and the same was found in similar experiments with electrolyte additions of 10
mM SmCl3 and 10 mM LaCl3 (curves not shown). For all three REM salt additions the
activation was limited to cathodic current densities lower than 100 A m-2, and it was also clear
from the polarization curves that hypochlorite reduction was hindered. The REM hydroxide
film could easily be seen by the naked eye on the electrode surface after experiments at 25°C,
but not after trials at 70°C. It was probably dissolved at the higher temperature as the current
was switched off.
-0.4
E/V vs Ag/AgCl
-0.6
70˚C
-0.8
50˚C
-1
-1.2
50˚C (without Y(III))
-1.4
25˚C
-1.6 1
10
10
2
i/Am
-2
10
3
Fig. 14: IR-corrected polarization curves obtained with a Fe-RDE in 0.5 M NaCl with 10 mM YCl3 at
varying temperatures. Electrode rotation rate 3000 rpm. (Paper 2)
Increasing the chloride concentration from 0.5 to 5 M NaCl had an effect on the Y(III)
activation at 25°C. It became less distinct and the potentials were unstable when recording
galvanostatic polarization curves, which were difficult to reproduce. When raising the
temperature to 70°C the activating effect of Y(III) addition vanished–see the curve for
5 M NaCl at 70°C in Fig. 15, showing an instant effect at ≈-100 A m-2 that disappeared at
lower current densities. To judge whether this behavior was related to the high chloride
concentration or to a high ionic strength in general, measurements were made in 5 M NaClO4
under the same conditions. In Fig. 15, the effect of high ionic strength and high chloride
concentration are separated at 70°C. The activation caused by Y(III) at -100 A m-2, that
disappeared when raising the chloride concentration to 5 M NaCl, was stable in 5 M NaClO4.
When at high concentration, the chloride ions therefore seem to be detrimental to the
catalysis. Note that the limiting current for hypochlorite reduction present in Fig. 15
disappeared in the presence of Y(III), indicating the presence of an inhibiting cathode film not
disturbed by a high chloride concentration. The negative effect above on the hydrogen
evolution catalysis may be explained by the formation of yttrium-chloride complexes that
lower the concentration of active Y(III) species. If the reactant of the catalyzed hydrogen
evolution is water complex bound to REM cations, the activation may arise from the Lewis
acidity of the cation weakening the O-H bond [51]. The energy needed to reduce such
24 complex bound water molecules is less than that needed to reduce free water. The
predominant Y(III) forms in 5 M Cl- at 25°C are yttrium chloride complexes (reaction 21 and
Fig. 16), which could affect the build-up and composition of the film. Furthermore, the Lewis
acidity of YCl2+ and other yttrium chloride complexes may be weaker than Y3+; water
molecules in the hydration shell can thus be less polarized by the electron-drawing cation
[52].
Y3+ + Cl-  YCl2+
Log K=-0.1
(21)
-0.4
5 M NaCl, pH=4.6
5 M NaClO4, pH=4.5
E/V vs Ag/AgCl
-0.6
5 M NaClO4, 10 mM YCl3, pH=4.5
-0.8
-1
5 M NaCl, 10 mM YCl3, pH=4.6
-1.2
-1.4
10
0
10
1
10
i/Am -2
2
10
3
Fig. 15: IR-corrected polarization curves obtained with a Fe-RDE at 70°C. Electrode rotation rate
3000 rpm. (Paper 2)
25 -1.0
-1.5
-2.0
Log Conc. / M
Y (O H )3
Y C l2 +
Y 3+
-2.5
-3.0
-3.5
Y O H 2+
-4.0
0
2
4
6
8
10
12
pH
14
Fig. 16: Log concentration vs. pH diagram generated with the Medusa Software [26]. The total
concentrations were set to 15 mM Y(III) and 5 M NaCl.
4.2.2Increasedhydrogenevolutionselectivitybytrivalentcationadditives
Hindering of hypochlorite reduction
As described earlier, current efficiency is important for the overall energy efficiency.
Potential sweeps were used to investigate the inhibition of hypochlorite reduction on iron
cathodes by REM salts (Fig. 17). With 15 mM hypochlorite present in the electrolyte, a
limiting current for hypochlorite reduction of -20 to -30 mA cm-2 was observed. In the
absence of hypochlorite, no limiting current was observed. The limiting current for
hypochlorite reduction was suppressed by the addition of Y(III), La(III) and Sm(III)
chlorides. The inhibition of hypochlorite reduction was probably caused by the formation of
REM hydroxide films. Care should be taken when comparing the inhibiting effects in Fig. 17,
since there were some bulk precipitation problems.
The use of stationary electrodes for current efficiency measurements showed that yttrium
hydroxide precipitated in the alkaline diffusion layer, rather than at the cathode surface. There
was a white cloud of precipitated particles around the electrode and the current efficiency for
HER did not increase after the addition. Good mass transport conditions are required to avoid
bulk precipitation of the trivalent metal hydroxides and instead form a cathode film.
26 No addition
0
10 mM SmCl3, 15 mM ClO-
i/mA/cm 2
-10
10 mM YCl3, 15 mM ClO-
-20
-30
10 mM LaCl3, 15 mM ClO-
15 mM ClO-
-40
-50
-1.4
-1.2
-1
-0.8
E/V vs Ag/AgCl
-0.6
-0.4
Fig. 17: Potential sweeps on an Fe-RDE at 3000 rpm in 5 M NaCl at 70°C, pH ≈5.4, swept
between -1.5 and -0.5 V at 50 mV s-1. The sweeps in the cathodic direction are not shown, but the
difference from the anodic sweeps was small. (Paper 2)
4.2.3Practicalimplicationsoftrivalentcationadditionwithemphasisonthechlorate
process
The optimal pH of the chlorate process is close to neutral pH, while for example Sc(III) and
Al(III) precipitate already at much lower pH values. Y(III), Sm(III) and La(III) will be very
close to precipitation around neutral pH. As observed in Paper 2, Y(III) precipitated already at
pH 4.8 in 550 g dm-3 NaClO3 at 70°C, which means that the Y(III) will be in precipitated
form at pH 6.5. It is possible that the trivalent cations also form precipitations with anions in
the electrolyte. In the chlorate process, the sodium chlorate is separated from the electrolyte
with crystallization at alkaline pH. If added to an industrial chlorate electrolysis process, the
trivalent cations would precipitate in this step and potentially contaminate the chlorate
product.
Furthermore, in Paper 2, it was found that high chloride concentrations form complexes with
Y(III) and probably also with other trivalent metal ions and thus decrease the activation of the
HER. High temperatures also decrease the degree of activation achieved by addition of
trivalent cations (e.g. Y(III)).
Given these problems with solubility, temperature and chlorides, trivalent metal ions cannot
be recommended as an additive to the chlorate process.
4.2.4Generaldiscussionoftrivalentcationadditives
Even if trivalent cations are not a realistic additive to the chlorate process electrolyte, there
are other applications where they may be suitable, for example MECs and water splitting solar
cells. The ability to activate the HER and inhibit side reactions at the same time is an
27 interesting combination. Instead of using a cell divided by for example a membrane, in a best
case scenario, the film on the cathode can inhibit the reduction of products formed at the
anode. In some applications, this is highly desirable, while in others it may have unwanted
effects. Electroplating in neutral electrolyte could for instance be problematic in the presence
of trivalent cations, since the deposition of metal might be suppressed while the HER may be
activated at the same time. Hydrogen evolution may also damage the deposit.
Another advantage of trivalent cations is that they are easily removed from the electrolyte by
increasing the pH. If a process step requires the electrolyte to be free of trivalent cations, the
pH is increased, the trivalent cations are removed by precipitation, the process step is
performed, the precipitate added again, the pH decreased and the trivalent cation can be
dissolved once again.
Buffers can be used instead of trivalent cations to activate the HER if there are no side
reactions on the cathode that need to be hindered. The main difference is that most buffers do
not form solid products after deprotonation and thus no hydroxide films that block the surface.
4.3Sodiummolybdateaselectrolyteadditive
4.3.1Activatedhydrogenevolutionbysodiummolybdate
Near neutral pH
Fig. 18 shows polarization curves for hydrogen evolution for molybdenum and titanium
rotating disc electrodes in 2 M NaCl at pH 6.5 in both the absence and presence of
4 mM Mo(VI). Without Mo(VI) in the electrolyte, both titanium and molybdenum were less
active than iron, with Tafel slopes of -340 mV dec-1 and -180 mV dec-1, respectively. When
reversing the polarization curves, titanium showed a more active electrode, indicating
non-stationary conditions. In the presence of Mo(VI), an in-situ activation of the cathodes
took place and the two substrates gave similar potentials for the (HER) and Tafel slopes in the
range of -170 to -180 mV dec-1. At technical current densities (between -2 and -3 kA m-2) the
in-situ activated electrodes were approximately as active as iron. Thus, as found earlier for
strongly alkaline conditions [37], the addition of Mo(VI) changes the electrode kinetics and
seemed to make the cathode potential relatively independent of the substrate material.
28 Fig. 18: IR-corrected polarization curves recorded in 2 M NaCl, pH≈6.5, 70°C and at a rotation rate of
5000 rpm. All polarization curves were recorded in the anodic direction. The electrodes were
pre-polarized at -3 kA m-2 for 15 min prior to the polarization curves. (Paper 4)
The Mo(VI) compounds in the electrolyte may be cathodically reduced to, for example,
Mo(IV) species. Current efficiency measurements were made to evaluate how large a
proportion of the applied current was related to this side reaction. A divided cell was used
[44] with stationary titanium electrodes and electrolytes of 2 M NaCl with and without,
respectively, 100 mM Mo(VI) at -3 kA m-2 and 70°C. Both an electrolyte of pH 6.5 and an
alkaline electrolyte containing 2 M NaCl, 1 M NaOH were investigated. These initial
measurements, made under vigorous magnetic stirring and hydrogen gas bubble generation,
showed current efficiencies for hydrogen production of 95-100% irrespective of whether
Mo(VI) was present or not and regardless of electrolyte pH. Thus, any reduction of Mospecies was only a minor part of the total applied current and the altered kinetics was not due
to a major change in electrode reaction from hydrogen evolution to Mo(VI) reduction. Also,
addition of Mo(VI) did not lead to a significant loss in current efficiency, which is important
for the energy efficiency of the process.
Since the buffering capacity in the presence of 4 mM Mo(VI) at neutral pH is quite low, we
also checked in Paper 4 that adding a phosphate buffer had no negative effect on the
activation by Mo(VI) on titanium and molybdenum RDEs at technical current densities.
To verify the role of Mo(VI) as forming catalytically active films, the following experiment
was performed: two titanium electrodes were pre-polarized for 30 min at -3 kA m-2 and
3000 rpm in 2 M NaCl at 70oC and pH 6.5 with and without the addition of 100 mM Mo(VI).
The electrodes were then rinsed in Milli-Q water and transferred to a Mo(VI) free electrolyte
of identical composition, and the electrode potentials were measured at -3 kA m-2. Electrodes
pre-polarized in a Mo(VI) containing electrolyte had about 200 mV lower overpotential for
hydrogen evolution compared to electrodes pre-polarized in the absence of Mo(VI). These
results indicate that the activation of hydrogen evolution relates to a catalytic film formed on
29 the cathode surface and not primarily to a mass transport controlled deprotonation of
electrolyte species, as found earlier in pH-neutral electrolytes for catalysis of the HER by for
example phosphate [17]. When 100 mM Mo(VI) was present in the electrolyte, the
overpotential for HER increased with time after the initial activation. At the end of the 30 min
period the activation was only 100 mV compared to titanium without addition of Mo(VI) to
the electrolyte (without IR-correction the electrode with molybdenum containing film became
with time even worse than the pure titanium). When electrodes with deposited molybdenum
containing films were transferred to a Mo(VI)-free electrolyte, the overpotential decreased
again. Obviously, there is an optimal Mo(VI) concentration to achieve the best activation of
the HER. In this study we have chosen to work mostly with additions of 4 mM Mo(VI), the
same concentration as Huot and Brossard [35] recommend for alkaline water electrolysis. In
Fig. 18 the observed activation of the HER by 400 mV on titanium by 4 mM Mo(VI) is
significantly higher than the activation of 100 mV with 100 mM Mo(VI) (not shown).
Alkaline electrolyte
As mentioned in the introduction, electrodeposited MoO2 is not stable at alkaline conditions
[40,42]. To investigate if this affects the activation for hydrogen evolution, cathodic
polarization curves were recorded in an electrolyte of 1 M NaOH. As can be seen in Fig. 19,
without additions the Tafel slopes were -160 mV dec-1 for titanium as cathode material
and -100 mV dec-1 for molybdenum. After polarization in the presence of 4 mM Mo(VI), the
Tafel slopes were the same for the two electrode substrates, -130 mV dec-1. Furthermore, the
electrodes were more active in the presence of Mo(VI) in the electrolyte and both titanium
and molybdenum substrates showed a similar potential for the HER. The activation thus
seemed to follow the same pattern irrespective of the electrolyte pH although different films
were formed; compare the results in Fig. 17 at near neutral pH with Fig. 18 at 1 M NaOH and
the data at 30 wt% KOH, 80°C, in ref [37]. Addition of Mo(VI) to the electrolyte forms films
on the cathode surfaces and the films rather than the substrate will determine the activity for
HER.
30 Fig. 19: IR-corrected polarization curves recorded in 1 M NaOH, 70°C at a rotation rate of 5000 rpm.
All polarization curves were recorded in the anodic direction. The electrodes were pre-polarized
at -3 kA m-2 for 15 min prior to the polarization curves. (Paper 4)
Surface analysis
Fig. 20, secondary electron mode SEM images are shown of titanium electrodes that had been
polarized with and without Mo(VI) in the electrolyte. A film, that contained molybdenum
according to EDS data, can be seen after polarization in the presence of Mo(VI) as well as
crystals of high molybdenum content and sodium chloride.
Fig. 20: SEM images on Ti-RDEs polarized for 30 min at -3 kA m-2 in 2 M NaCl, pH≈6.5 and 70°C
with and without 4 mM Mo(VI). The rotation rate during polarization was 3000 rpm. Composition
was analyzed with EDS in atomic % and the Mo/(Mo+Ti) ratio will be given in the following points:
a.) 23%, b) 26%, c.) 6%, d.) 9% (Paper 4)
Further surface analyses were performed on titanium cathode samples that had been polarized
at -3 kA m-2 for 30 minutes in either 2 M NaCl at near neutral pH or in 1 M NaOH, both with
addition of 4 mM Mo(VI). Analyses by XRF showed that there was significantly more
31 molybdenum on electrodes polarized in neutral media than in alkaline electrolyte (Fig. 21).
The Mo K-L3 and K-M3 peaks were in good agreement with data available in literature [53].
Fig. 21: XRF-spectra for Ti-RDEs polarized for 30 min at -3 kA m-2, 70°C and at a rotation rate of
3000 rpm. The pH was adjusted to 6.5 for the electrolyte containing 2 M NaCl, 4 mM Mo(VI).
(Paper 4)
Data from XPS measurements also showed that there was more molybdenum on electrodes
polarized in neutral than alkaline electrolyte. With XPS it was possible to detect molybdenum
species in the outermost surface layer even for electrodes polarized in alkaline electrolyte.
Considering that Mo(VI) addition can activate the HER also under alkaline conditions (Fig.
19), this agrees well with the fact that molybdenum could be detected with XPS, as it is a very
surface-sensitive technique.
4.3.2Increasedhydrogenevolutionselectivitybysodiummolybdateadditive
Hindering of hypochlorite reduction by Mo(VI) addition
Potential sweep experiments on a Ti-RDE with and without additions of Mo(VI) and
hypochlorite to the electrolyte are shown in Fig. 22. The sweeps were made in anodic
direction after 5 minutes of pre-polarization at -1.5 V vs. Ag/AgCl, which enabled the
formation of Mo-containing films on the cathode whenever the electrolyte contained Mo(VI).
In the presence of 15 mM hypochlorite, the dotted line, a cathodic limiting current for
hypochlorite reduction of about 350 A m-2 can be seen. This limiting current was suppressed
by the addition of Mo(VI), and the higher the Mo(VI) concentration the more efficient the
hindering effect – see curves with addition of 1, 10 and 100 mM Mo(VI), respectively.
Looking at the cathodic end of the graph an activation of hydrogen evolution is seen for the
cases of 1 and 10 mM Mo(VI). As high an Mo(VI) level as 100 mM did not give a similar
decrease in overpotential, probably caused by resistive properties of the Mo(VI) film formed
[Paper 4].
32 Fig. 22: Potential sweeps from -1.5 V to 0 V vs. Ag/AgCl at a sweep rate of 50 mV/s in 5 M NaCl,
70°C, pH 6.5 with a Ti-RDE at a rotation rate of 3000 rpm. Prior to the sweeps the electrode had been
polarized at -1.5 V vs. Ag/AgCl for 5 minutes. (Paper 5)
Fig. 23 shows potential sweeps on slightly corroded iron with the electrode pre-polarized as
above. Again, a limiting current for hypochlorite reduction is visible at -300 to -400 A m-2,
and addition of 100 mM Mo(VI) only partly suppressed this current. A comparison with the
result in Fig. 22 indicates that it is easier to form a hindering film on a titanium substrate than
on corroded iron. The high currents appearing at potentials > -0.6 V vs. Ag/AgCl in the
Mo(VI) free case likely relate to oxidation of the iron electrode, which is hindered by the
presence of a molybdenum containing film. Similar to the case of titanium cathodes, the
addition of as high a concentration as 100 mM Mo(VI) did not result in activation of the HER
and instead deactivated the iron cathode.
33 Fig. 23: Potential sweeps from -1.5 V to 0 V vs. Ag/AgCl at a sweep rate of 50 mV s-1 in 5 M NaCl,
pH 6.5, 70°C with a corroded Fe-RDE at a rotation rate of 3000 rpm. Prior to the sweeps the electrode
had been polarized at -1.5 V vs. Ag/AgCl for 5 minutes. (Paper 5)
Hindering of hypochlorite reduction by Cr(VI) addition
Potential sweep experiments were made on titanium in electrolytes with varying Cr(VI)
concentrations (Fig. 24) to study the effect of low Cr(VI) concentrations in chlorate
electrolyte. In the 5 M NaCl electrolyte with 15 mM hypochlorite there was a limiting current
for hypochlorite reduction. This electrolyte was Cr(VI)-free and there was no inhibition of
hypochlorite reduction. A similar sweep was made in chlorate electrolyte made from
commercial grade NaClO3 and p.a. grade NaCl, and showed that hypochlorite reduction was
inhibited in the presence of 550 g dm-3 NaClO3. From ICP analyses of the chlorate crystals it
was calculated that this electrolyte also contained about 20 μM Cr(VI) and this inhibition was
reproduced by adding 20 μM Cr(VI) to 5 M NaCl, see Fig. 24. This inhibition of hypochlorite
reduction on titanium in chlorate electrolyte without added Cr(VI) was thus caused by the
Cr(VI) traces in non-recrystallized chlorate salt. Experiments were also performed with an
electrolyte made from NaClO3 that had been recrystallized by means of cooling crystallization
in the lab. For electrolytes with recrystallized chlorate, the limiting current for hypochlorite
reduction was higher, but the chromate effect was still present. This makes it difficult to
evaluate Cr(VI) free alternatives using chlorate salt produced from a conventional industrial
process with Cr(VI) containing electrolyte and the present study thus considers only
hypochlorite reduction (reaction 18) and not chlorate reduction (reaction 17).
34 Fig. 24: Potential sweeps (anodic direction) using a Ti-RDE at 3000 rpm and 70°C in electrolytes
containing 15 mM hypochlorite performed with a sweep rate of 50 mV s-1, Pre-polarization at -1.5 V
vs. Ag/AgCl for 5 minutes. (Paper 5)
The effect of low levels of Cr(VI) was also studied in CE measurements on a steel electrode,
see Fig. 25 where, for comparison, a curve representing Mo(VI) additions has been inlaid.
The addition of 2.7 M Cr(VI) increased the CE, as measured after 20 minutes of electrolysis,
from about 80% to over 94%. This can be compared to 80 mM Mo(VI), which only increased
the CE to about 83% despite being present in a concentration about 30,000 times higher.
Increasing the Cr(VI) level to 27 M increased the CE to 97% after 20 minutes of
electrolysis. With the lower Cr(VI) concentration it took over 10 minutes to reach steady state
CE values, whereas for 27 M Cr(VI) a high CE was reached instantaneously (in less than
1 min). Adding 80 mM Mo(VI) to the Cr(VI)-containing electrolyte did not have a negative
effect on the CE, but instead a small positive effect. Note that the high CE values obtained in
these experiments relate to a steel cathode that had been polished prior to the experiments and
submerged into the electrolyte under cathodic polarization to avoid steel corrosion. If the
electrode had been allowed to corrode on open circuit in hypochlorite containing electrolyte,
the CE values would have been lower [54] and probably the very low Cr(VI) levels in the
present study would have had a negligible effect on the CE. For corroding steel, higher
chromate levels are therefore needed [54], but if a smooth and relatively stable substrate
material is used the Cr(VI) level in the chlorate process may be lowered substantially while
maintaining a high current efficiency. This seems true also for Mo(VI) addition and therefore
it will be difficult to replace Cr(VI) in the existing process without replacing the steel
cathodes with a more dimensionally stable material.
35 Fig. 25: CE measurement for HER on a stationary steel electrode in 1.6 M NaCl at 70°C with an initial
hypochlorite concentration of approximately 80 mM and initial pH of 6.5. (Paper 5)
4.3.3Effectsontheanodebysodiummolybdateaddition
To study anode effects of Mo(VI), RuO2/TiO2 electrodes produced by spin-coating and
shaped into RDEs were polarized at 3 kA m-2 for 30 minutes with and without 40 mM Mo(VI)
in the electrolyte. The addition of Mo(VI) resulted in an increase in anode potential of around
40 mV, see Fig. 26. Using SEM, no clear morphology change due to film formation or
deposits could be seen on the RuO2/TiO2 surfaces, although Mo/Na ratios 2-10 times larger
than expected from electrolyte composition were measured with EDS. XRF measurements
also indicated an increased molybdenum content on the surface. Furthermore, predominance
diagrams show MoO3·2H2O(s) as the dominant phase in 10 mM Mo(VI) solution if the
conditions are acidic and oxidizing, such as close to the anode [42]. It is thus likely that the
increased oxygen is caused by precipitation or adsorption of Mo-containing species on the
anode surface.
36 Fig. 26: Galvanostatic polarizations of DSA‐RDEs (spincoated 3‐layers) at 3 kA m‐2 in 2 M NaCl, pH 7, 70°C at a rotation rate of 3000 rpm. (Paper 5) 4.3.4Practicalimplicationsofsodiummolybdateadditionwithemphasisonthechlorate
process
Sodium molybdate has the potential to become an important additive to the chlorate process.
It can activate the HER (on electrode substrates with poor activity for the HER) and help
hinder side reactions and gives a freedom of choice between different cathode substrates
without sacrificing the activity of the HER.
The cathode overpotential increases if the Mo(VI) concentration is raised over a certain
threshold level. For example, the overpotential was higher in the presence of 100 mM Mo(VI)
than in the presence of 4 or 10 mM Mo(VI). The overpotential of the anode is higher in the
presence than in the absence of 40 mM Mo(VI) and in the literature an increased oxygen level
in the cell gas has been measured in the presence of 39 mM Mo(VI) [32]. These effects can
for example be caused by adsorption or deposition of a film on the anode. Increased cathode
potential, increased anode potential, and increased oxygen level in the presence of high
Mo(VI) concentrations give very strong motivations to use a low level of Mo(VI) in the
process. Even if the partial current for the side reaction of Mo(VI) reduction appeared to be
quite low even for 100 mM Mo(VI), the current efficiency of the HER would probably be
even higher if a low Mo(VI) level was used, thus giving another motivation to use a low
Mo(VI) level.
A low concentration of Mo(VI) alone will not be sufficient to achieve a high current
efficiency in the chlorate process. Some other additive therefore also has to be used, for
example low levels of Cr(VI). If both the Cr(VI) and Mo(VI) concentrations are low, then a
third additive such as phosphate would be required to help buffer the pH of the process. It is a
significant gain for worker safety and the environment if the Cr(VI) concentration can be
decreased from what is used today.
37 4.4Comparingtrivalent,sodiummolybdate,buffersandCr(VI)electrolyte
additives
Trivalent cations investigated in this study are in a constant oxidation state as M(III). The ions
generally form positive M(H2O)x3+ complexes in acidic electrolytes and the hydroxide
M(OH)3·xH2O in alkaline electrolytes. Molybdate (MoO42-) and chromate (CrO42-) on the
other hand are in the hexavalent oxidation state and in the form of oxyanions. These
oxyanions buffer the pH around neutral [32], which would help to control the pH in for
example the chlorate process.
Trivalent cations require an acidic to neutral pH to be dissolved. This limits the pH range
where trivalent cations can be used to activate the HER. The ligand water of the trivalent
cations can be electrochemically deprotonated more easily than ordinary water molecules due
to a lower pKa value, although hydronium ions with a pKa value of -1.74 [13] are even more
easily deprotonated. The pH limits for activation by trivalent cations are therefore set between
the pH where they precipitate as hydroxides and the pH where the hydronium ions in the
absence of trivalent cations become abundant enough to sustain the same activity. Buffers can
give a similar activation to trivalent cations, but do not deposit as a hydroxide film on the
cathode surface. Either the buffer is electrochemically deprotonated or the buffer increases the
hydronium ion concentration at the cathode surface and thus decreases the concentration
polarization of hydronium ions.
Achieving an activation of the HER by trivalent cations or buffer additives at technical
current densities (-2 to -3 kA m-2) would require a good mass transport. Additives that form
catalytically active films on the other hand can be quite independent of mass transport
conditions, since often a few monolayers of deposit is enough to produce an effect.
Mo(VI) is dissolved in the electrolyte over a wide pH range, all pH values over 3 for 10 mM
Mo(VI) [42]. Mo(VI) is reduced to lower oxidation states on the cathode and forms
molybdenum oxides or alloys with other metals. Mo(VI) added to electrolytes of a wide pH
range therefore activates the HER by forming catalytically active films.
Cr(VI) added to the electrolyte is reduced to very thin (a few nm) films of Cr(OH)3·H2O
which not only increases the selectivity of the HER, but also prevents its own growth [22].
Films formed from Cr(VI) therefore do not grow too thick. Mo(VI), on the other hand, when
added to a pH neutral electrolyte, forms thicker cathode films which are easily visible with the
naked eye, often as a black deposit of molybdenum oxides. Films formed from Mo(VI)
additive are far less efficient at suppressing hypochlorite reduction than films formed from
Cr(VI). Films formed from high levels of Mo(VI) also appear to grow too thick and give
higher resistance for the HER than films formed from low Mo(VI) levels. The films formed in
the presence of Mo(VI) and Cr(VI) require a Faradaic reaction and it is therefore easier to
limit the reactions to the cathode surface. In other cases, such as in the case with trivalent
cation additions, pH dependent bulk precipitation may occur instead of the formation of
cathode films.
38 5.Conclusions
The objective of this work was to better understand activated hydrogen evolution reaction
from pH-neutral electrolytes. Special emphasis was laid on considering the potential
application in the chlorate process.
5.1TrivalentcationsasadditivetopHneutralelectrolytes
Addition of trivalent cations to weakly acidic or pH neutral electrolytes can activate the
electrolytic HER. This activation increases with the acidity of the metal ion. Acidic metal ions
polarize the ligand water and thereby decrease the pKa value. A metal hydroxide film is
precipitated at the cathode surface during hydrogen evolution. With time, this film may
continue to build up, block the surface, and thereby deactivate the HER.
The hydroxide films formed can inhibit parasitic side reactions to HER. This increases the
current efficiency for HER, which lowers the electric energy consumption. If the alkaline
diffusion layer at the cathode is too thick, the ions will precipitate in the electrolyte before
reaching the surface. The ability to form hydroxide films is thus dependent on the mass
transport conditions and the rate of hydroxide production.
Trivalent cations are not suitable additives in the chlorate process. They generally precipitate
in chlorate electrolyte at pH-values lower than what is optimal for the process. Chloride ions
present in the chlorate process may have a negative effect on the activation if chloride
complexes form with the trivalent cations. The activating effect of trivalent cations also
appears to decrease with increasing temperature, which makes the additive less relevant for
electrolysis at the elevated temperatures in the chlorate process. Trivalent cations can be a
promising additive to activate the HER at room temperature and chloride levels below 0.5 M.
It may be possible to use in applications such as electrochemical waste water treatment and
MEC, although further studies are needed.
5.2SodiummolybdateasadditivetopHneutralelectrolytes
In molybdate ions (MoO42-), the molybdenum is in the hexavalent oxidation state. Addition of
molybdate to neutral and alkaline electrolytes activates the electrolytic HER (on electrode
substrates with poor activity for the HER). The added molybdate ions are reduced on the
cathode forming for example films of oxides or alloys. These films give different electrode
substrates a similar activity for HER. Films formed from neutral electrolyte in this study
mainly consisted of molybdenum oxides and was thicker than films formed from alkaline
electrolyte. The film also maintains its catalytic ability for the HER when the electrode is
transferred to a molybdate free electrolyte.
The activation for neutral electrolytes was larger with 4 mM molybdate as electrolyte additive
than with 100 mM. Large amounts of molybdate appear to be detrimental to the activation.
Molybdate also increased the overpotential on the anode due to either adsorption or
precipitation of film.
Side reactions to the HER can be inhibited by molybdate additive. The additive is more
efficient when added to neutral than to alkaline electrolytes. In alkaline electrolytes, films of
39 molybdenum oxides are not stable and thus less efficient at hindering for example
hypochlorite reduction. For inhibiting side reactions, molybdate is far less efficient than
dichromate. Considering that even recrystallized chlorate from industrial production contains
enough Cr(VI) to hinder hypochlorite reduction on titanium, care should be taken when
interpreting reports that claim to be performed Cr(VI)-free and still use industrial chlorate.
Molybdate can potentially be used as an additive in the chlorate process. The concentration of
molybdate should be low since high molybdate concentrations have negative effects on
cathode (100 mM MoO42-) and anode overpotential (40 mM MoO42-). A low molybdenum
concentration present in the chlorate electrolyte would give greater freedom in the choice of
cathode material. The substrate material is not required to be especially active for HER itself
since the molybdenum-containing film formed will activate the HER. A stable and
inexpensive material with good electric conductivity and good mechanic strength would be
interesting, although long-term effects such as the deposition of impurities, the effect on the
anodes, and oxygen levels in cell gas need to be carefully evaluated.
Mo(VI) cannot yet be considered a replacement for Cr(VI) in the chlorate process. The
molybdate additive can to some extent hinder the hypochlorite reduction, although high
enough molybdate concentrations increase anode and cathode potentials. Therefore the
molybdate additive needs to be complemented with either Cr(VI) or some other additive to
yield acceptable current efficiencies. To fully replace Cr(VI) or significantly lower the
concentration, a buffer agent such as phosphate will be needed.
40 6.Futurework
More work is required to investigate trivalent cations and molybdate as additives in different
processes. It would be interesting to see if these additives work in MECs, water splitting solar
cells, and on hydrogen-evolving counter-electrodes in electrochemical water purification.
Relevant long-term studies are difficult to perform in lab-scale experiments and would
probably be best performed in a pilot scale plant. Pilot scale experiments can also give some
indication of unexpected problems that may arise, such as chlorate reduction.
To study the combination of molybdate and trivalent cations would also be very interesting,
but perhaps not relevant for application in the chlorate process.
I think a promising way to move forward in the research on the chlorate process would be to
test a new concept in lab-scale experiments and a pilot scale plant: A stable and smooth
cathode material, used in combination with in-situ additives to the electrolyte. Low levels of
Mo(VI) should be added for the activity and low levels of Cr(VI) for the current efficiency of
the HER. One or several other additives should also be added to replace some of the functions
of the higher Cr(VI) levels used today. Perhaps it is possible to find individual additives that
replaces individual effects of the Cr(VI).
Additives that enhance the effects of Mo(VI) and Cr(VI) would also be interesting.
7.Acknowledgements
Eka Chemicals, Permascand and the Swedish Energy Agency are acknowledged for their
financial support.
I would like to thank Ann Cornell, Göran Lindbergh and Joakim Bäckström for being my
supervisors.
I thank Linda Nylén, Christine Hummelgård and everyone else with whom I have
collaborated.
I thank all my colleagues at the Division of Applied Electrochemistry for all these nice years
during my PhD studies.
41 42 8.References
1. R. F. Savinell, Our Featured Division:Industrial Electrolysis and Electrochemical
Engineering Division, ECS interface Fall (1998), p. 30
2. S. Trasatti, “Electrocatalysis of Hydrogen Evolution: Progress in Cathode
Activation” in H. Gerischer and C. W. Tobias, Advances in Electrochemical Science
and Engineering, Vol 2 (1991) VCH Publishers Inc., New York, USA
3. G. Aylward and T. Findlay, SI Chemical Data 4th edition, (1998), Jacaranda Wiley
Ltd, Milton, Australia
4. P. Marcus, Corrosion Mechanisms in Theory and Practice, 2nd edition, (2002) Marcel
Dekker Inc., New York, USA
5. D. Simonsson, Grundläggande och tillämpad elektrokemi, (1987) Institutionen för
Teknisk Elektrokemi och Korrosionslära, KTH, Stockholm, Sweden
6. V. S. Bagotsky, Fundamentals of Electrochemistry, 2nd edition, (2006) John Wiley &
Sons, Inc., Hoboken, New Jersey, USA
7. V. Artero, M. Chavarot-Kerlidou and M. Fontecave, Splitting water with cobalt,
Angewandte Chemie 50 (2011) 7238-7266
8. A. W. Jeramiasse, J. Bergsma, J. M. Kleijn, M. Saakes, C. J. N. Buisman, M. C. Stuart
and H. V.M. Hamelers, Performance of metal alloys as hydrogen evolution reaction
catalysts in a microbial electrolysis cell, International Journal of Hydrogen Energy 36
(2011) 10482-10489
9. A. W. Jeramiasse, H. V.M. Hamelers, J. M. Kleijn and C. J. N. Buisman, Use of
biocompatible buffers to reduce the concentration overpotential for hydrogen
evolution, Environmental Science and Technology 43 (2009) 6882-6887
10. M. D. Merril and B. E. Logan, Electrolyte effects on hydrogen evolution and solution
resistance in microbial electrolysis cells, Journal of Power Sources 19 (2009) 203-208
11. L. Nylén and A. Cornell, Effects of electrolyte parameters on the Iron/Steel cathode
potential in the chlorate process, Journal of Applied Electrochemistry 39 (2009) 71-81
12. T. Hurlén, S. Gunvaldsen and F. Blaker, Effects of buffers on hydrogen evolution at
iron electrodes, Electrochimica Acta 29 (1984) 1163-1164
13. R. Starkey, J. Norman and M. Hintze, Who Knows the Ka values of Water and the
Hydronium ion, Journal of Chemical Education 63 (1986) 473-474
14. L. De Silva Menoz, A. Bergel, D. Feron and R. Basséguy, Hydrogen production by
electrolysis of a phosphate solution on a stainless steel cathode, International Journal
of Hydrogen Energy 35 (2010) 8561-8568
15. V. Marinovic and A. R. Despic, Cathodic hydrogen evolution from aqueous solutions
of acetic acid, Russian Journal of Electrochemistry 40 (2004) 995-999
16. S. Da Silva, R. Basséguy and A. Bergel, Electrochemical deprotonation of phosphate
on stainless steel, Electrochimica Acta 49 (2004) 4553-4561
17. V. Marinovic and A. Despic, Phosphoric acid as a source of hydrogen in cathodic
hydrogen evolution, Journal of the Serbian Chemical Society, 63 (1998) 545-553
18. V. Marinovic and A. R. Despic, Hydrogen evolution from solutions of citric acids,
Journal of Electroanalytical Chemistry 431 (1997) 127-132
43 19. V. Marinovic and A. Despic, Pyrophosphoric acid as a source of hydrogen in cathodic
hydrogen evolution on silver, Electrochimica Acta 44 (1999) 4073-4077
20. T. N. Andersen, B. S. Dandapani and J. M. Berry, Hydrogen evolution studies in
neutral media, Journal of Electroanalytical 357 (1993) 77-89
21. J. E. Colman, Electrolytic production of sodium chlorate, American Institute of
Chemical Engineers Symposium Series 77 (1981) 244-263
22. G. Lindbergh and D. Simonsson, Effects of chromate addition on cathodic reduction
of hypochlorite in hydroxide and chlorate solutions, Journal of the Electrochemical
Society 137 (1990) 3094-3099
23. A. Ahlberg Tidblad and G. Lindbergh, Surface analysis with ESCA and GD-OES of
the film formed by cathodic reduction of chromate, Electrochimica Acta 36 (1991)
1605-1610
24. G. Lindbergh and D. Simonsson, Inhibition of cathode reactions in sodium hydroxide
solution containing chromate, Electrochimica Acta 36 (1991) 1985-1994
25. K. Viswanathan and B. V. Tilak, Chemical, Electrochemical and Technological
Aspects of Sodium Chlorate Manufacture, Journal of the Electrochemical Society 131
(1984) 1551-1559
26. M. M. Jaksic, J. Brun, B. Johansen and R. Tunold, The Rowland or EDTA effect on
electrochemical behavior of transition metals and in electrocatalysis for the hydrogen
evolution reaction. The Rowland effect on some non-noble non-valve transition
metals, Russian Journal of Electrochemistry 31 (1995) 1187-1202
27. Medusa software, 32-bit version, updated 6th December 2010,
http://www.kemi.kth.se/medusa/
28. M. Tran, C. Fiaud and E. M. M. Sutter, Cathodic reactions on steel in aqueous
solutions containing Y(III), Journal of the Electrochemical Society 153 (2006)
B83-B89
29. M. Tran, D. Mohammedi, C. Fiaud and E. M. M. Sutter, Corrosion behaviour of steel
in the presence of Y(III) salts: kinetic and mechanistic studies, Corrosion Science 48
(2006) 5257-5273
30. M. Tran, P. Dubot and E. M. M Sutter, Activation of water in the presence of REM
salts in aqueous solution, International Journal of Hydrogen Energy 33 (2008)
937-944
31. C. T. Hsu and S. K. Yen, Electrolytic Y2O3 coating on IN617 Superalloy, Journal of
the Electrochemical Society 152 (2005) C813-C816
32. M. Li, Z. Twardowski, F. Mok and N. Tam, Sodium molybdate – a possible alternate
additive for sodium dichromate in the electrolytic production of sodium chlorate,
Journal of Applied Electrochemistry 37 (2007) 499-504
33. M. Rosvall, K. Hedenstedt, A. Sellin, J. Gustavsson and A. Cornell, Activation of
cathode, (Akzo Nobel Chemicals International B. V.), WO 2010/130546 A1 (2010)
34. J. Y. Huot and L. Brossard, Time dependence of the hydrogen discharge at 70°C on
nickel cathodes, International Journal of Hydrogen Energy 12 (1987) 821-830
35. J. Y. Huot and L. Brossard, Activation in situ de la cathode de nickel par le molybdate
de sodium, Surface and Coatings Technology 34 (1988) 373-382
44 36. J. Y. Huot and L. Brossard, In situ activation of cobalt cathodes in alkaline water
electrolysis, Journal of Applied Electrochemistry 18 (1988) 815-822
37. J. Y. Huot and L. Brossard, In situ activation by sodium molybdate on various metallic
substrates during alkaline water electrolysis, International Journal of Hydrogen Energy
14 (1989) 229-232
38. J. Y. Huot and L. Brossard, In situ activation of nickel cathodes by sodium molybdate
during alkaline water electrolysis at constant current, Journal of Applied
Electrochemistry 20 (1990) 281-288
39. L. Brossard and J. Y. Huot, In situ activation of cathodes during alkaline water
electrolysis by dissolved iron and molybdenum species, Journal of Applied
Electrochemistry 21 (1991) 508-515
40. J. Presuel-Moreno, M. A. Jakab and J. R. Scully, Inhibition of the oxygen reduction on
copper with cobalt, cerium and molybdate ions, Journal of the Electrochemical Society
152 (2005) B376-387
41. A. Brenner, Electrodeposition of Alloys, 1-2, Academic Press Inc., New York, USA,
1963
42. P. Wang, L. L. Wilson, D. J. Wesolowski, J. Rosenqvist and A. Anderko, Solution
chemistry of Mo(III) and Mo(IV): Thermodynamic foundation for modeling localized
corrosion, Corrosion Science 52 (2010) 1625-1634
43. A. Cornell, B. Håkansson and G. Lindbergh, Ruthenium based DSA in chlorate
electrolysis-critical anode potential and reaction kinetics, Electrochimica Acta 48
(2003) 473-481
44. A. Cornell, G. Lindbergh and D. Simonsson, The effect of addition of chromate on the
hydrogen evolution reaction and on iron oxidation in hydroxide and chlorate solutions,
Electrochimica Acta 37 (1992) 1873-1881
45. M. Stern, The Electrochemical Behavior, Including Hydrogen Overvoltage, of Iron in
Acid Environments, Journal of the Electrochemical Society 102 (1955) 609-616
46. A. C. Riddiford, In Advances in Electrochemistry and Electrochemical Engineering,
Vol 4, p. 61, John Wiley & Sons, New York, USA, 1966
47. I. Zhitomirsky and A. Petric, Electrochemical deposition of yttrium oxide, Journal of
Materials Chemistry 10 (2000) 1215-1218
48. M. Heyrovsky, Polarographic wave of aluminium, Collection of Czechoslovak
Chemical Communications 25 (1960) 3129-3135
49. C. F. Baes and R. E. Mesmer, The hydrolysis of cations (1976) Wiley, New York,
USA
50. S. Vasudevan, Studies relating to cathodic reactions in neutral chloride solutions used
in chlorate process, Indian Engineering Chemistry Research 47 (2008) 5742-5745
51. G. Wulfsberg, Inorganic chemistry, p. 56, (2000) University Science Books, Sausalito,
California, USA
52. R. M. Smith and A. E. Marell, Critical stability constants, Vol 4: inorganic complexes
p. 104, (1976) Plenum Press, New York, USA
53. J. A. Bearden, X-Ray Wavelengths, Reviews of Modern Physics 39 (1967) 78-124
54. J. Wulff and A. Cornell, Cathodic current efficiency in the chlorate process, Journal of
Applied Electrochemistry 37 (2007) 181-186
45